Study Limitations

Our study has limitations. We evaluated a commercially insured population; rates of dispense as written requests and prescription reversals may differ for uninsured patients. Our measure of reversal was linked to the specific prescription that was adjudicated by the pharmacy. Some of the unfilled prescriptions may not represent clinically significant medication non-adherence, because patients may have requested new prescriptions for different medications to treat the same condition. However, previous studies indicate that patients who are initially prescribed branded medications are less likely to subsequently adhere to any medication in the class when compared with patients prescribed generics. We recruited during a 1-month period in the winter. We did not account for seasonality; patient medication use and prescription requests may vary by season. We also are unsure of the extent of misclassification in this data set, because pharmacies may not accurately capture all patient dispense as written requests in administrative data sets, which may have led to conservative estimates of dispense as written rates.

Conclusions

Overall, we found that both patients and physicians commonly make dispense as written requests, totaling approximately 5% of all prescriptions. Advocates of dispense as written may argue that providing physicians and patients with greater discretion offers greater choice, opportunities for communication, and adherence to therapy. However, our results indicate that dispense as written requests are associated with excess costs, and that patients are less likely to fill prescriptions with dispense as written designations. Some private health plans have implemented financial penalties to reduce the rates of dispense as written designation. The cost savings and clinical effects of these policies should be studied to better understand what policies best encourage cost-effective medication use and adherence to chronic therapy.